Simulated Intelligent Robot Tracking Agent

Introduction

Simulated Intelligent Robot Tracking Agent: in course project, I developed a naive intelligent agent to predict the future trajectory of a Nano robot’s dynamic moving position; evaluated multiple training algorithms in Bayesian probabilistic model, linear-Gaussian model (Kalman Filters), sequential Monte Carlo simulation (particle filters), residual learning model; reduced video data dimensionality by PCA; tuned residual neural network hyperparameters and applied bootstrap aggregation with multiple residual neural networks [Github][PDF].

Results